Skip to main content

Full text of "Through the Great Canyon of the Euphrates River"

See other formats


Geographical Journal 



VOL. XX. — July to December, 1902. 


EDWARD STANFORD, 12, 13 and 14, Long Acre, W.C. 


Digitized by 




Through the Great Canon of the Euphrates River. By Ellsworth Huntington 

(with Map and 9 Illustrations) 175 

Report on the Identification of the Bay of San Felipe and Santiago visited 

by Quiros in 1606 201 

Note on the Identification of La Sagittaria of Quiros. By Rear-Admiral 

Sir W. J. L. WhartoD, K.C.B., P.B.S. 207 

The National Antarctic Expedition. The Departure of the Morniny 

(with Illustration and Portrait) 209 

The Russian Polar Expedition. Summary Report for 1901. By Baron 

Ed. Toll 216 

Reviews : — 

America— Central America 220 

Oceanography — The Voyage of the Gauss 222 

The Monthly Record 224 

Correspondence 235 

(Geographical Literature of the Month 237 

New Maps 252 

No. 3. September. 

The Geographical Conditions determining History and Religion in Asia 

Minor. By Prof. W. M. Ramsay (with Map) 1257 

The Caura Affluent of the Orinoco. By E. Andre* (with Sketch-map) 283 
Summary of the Results of Dr. Sven Hedin's Latest Journey in Central 
Asia (1899-1902). By Dr. Sven Hedin (with Sketch-map, 4 Illustra- 
tions, and Diagram) 307 

The International Council for the Study of the Sea 316 

The "Sudd" of the White Nile. By Edward S. Crispin, m.b.c.s., l.r.c.p. 

(with 9 Illustrations) 318 

Geographical Results of the Explorations of the French " White Fathers " 

in North-Eastern Rhodesia. By Henri Maitre (with Sketch-map) .. 324 
Reviews : — 

Africa— Uganda 328 

America — Charlevoix's 'New France 1 331 

Polar Regions— Dr. Nansen's Third Volume of Scientific Results .. 332 

The Monthly Record 334 

Obituary 347 

Geographical Literature of the Month 349 

New Maps 366 


Sketch-map of Venezuela 285 

Sketch-map showing Dr. Sven Hedin's Routes in Central Asia 308 

Sketch-map of Journeys of the French •' White Fathers " in Lobemba and 

Lobisa 325 

Map of Asia Minor 372 

No. 4. October. 

From the Somali Coast through Southern Ethiopia to the Sudan. By Oscar 

Neumann (with 11 Illustrations and Map) 37o 

Digitized by 



other room some of his photographs, which were examples of the way in which, 
by a knowledge of the formation of ripples in sand, one would he able to tell the 
direction of currents in mouths of rivers and estuaries. So I think we can con- 
gratulate Dr. Yaughan Cornish, not only on the great scientific value of his 
investigations, but also on the practical use to which many of them may be turned. 
I now ask the meeting to pass a unanimous vote of thanks to Dr. Yaughan Cornish 
for his valuable paper and for the beautiful way in which he has illustrated it by 
his photographs. 



A reference to the map of Asia Minor shows that the Euphrates river 
is formed by the union of two great branches. These for several 
hundred miles flow west- south-west in nearly parallel longitudinal 
valleys bounded by high ridges of mountains. The north-western 
branch, or Kara Su, although it is smaller than the other, is generally 
called the Fiat, or Euphrates. Near Egin the mountains that bound 
its valley oome together, and it is obliged to turn abruptly at right 
angles toward the south through a tremendous gorge, which it has cut 
directly across the southern range of the anti-Taurus mountains. 
Twenty miles below the gorge it empties into the larger branch, the 
Murad Su. The latter flows from north of Van nearly straight west- 
south-west to this point, part of the way through an unexplored canon, 
said to be longitudinal and impassable. Below the junction of the 
Murad Su and Kara Su the Euphrates river makes a great bend to the 
south, and later to the east, and then, again turning south, zigzags 
through the Taurus mountains in a deep gorge. It is this great 
bend and the lower part of the two branches which I propose to 

Before proceeding to that, however, it will be well to get a general idea 
of the physical features of the surrounding country. We will examine 
these in geographical, order, beginning on the south with the Taurus range. 
This rises here to an average height of 7000 feet, the lowest passes, except 
where the Euphrates and Tigris rivers cut through the range, having 
an elevation of about 5000 feet. The core of the mountains consists of 
a very old trap, the oldest formation of the region. This is broken in 
many places by eruptive granites and porphyries, which are older than 
the oldest stratified deposits. These crystallines are especially abundant 
on the north side of the mountains. On the weathered surface of these 
igneous formations are the remnants of a complex series of strata with 
several unconformities marking intervals of erosion so extensive that 

♦ Map, p. 177. 

Digitized by 



only the upper series of strata remain in large quantities. These are 
Cretaceous shale and limestone containing, among other fossils, numerous 
hippurites and nutnmulites. In post-Cretaceous times extensive 
elevation and mountain-building took place accompanied by considerable 
faulting. Probably it was at this time that the shale was locally much 
metamorphosed. The Cretaceous strata, rarely exceeding 1000 feet in 
thickness, form the tops of the majority of the mountains. Overlying 
the uplifted and highly eroded surface of this formation are numerous 
recent lava-flows, chiefly basaltic, which evidently took place after the 
country had assumed approximately its present topography. In several 
places there are cones of ash, and south of Lake Gyuljuk there is a 
crater enclosing a pond. 

This lake lies in a longitudinal valley between the northern 
ridges of the Taurus, in the neck of the great bend of the Euphrates 
river. Its position in respect to the latter is so remarkable that it 
deserves notice. The map shows that below Palu the river flows south- 
west directly toward the lake and in the same longitudinal valley. 
Fifteen miles from Gyuljuk it is turned west and north around the 
great bend already mentioned. After a course of 160 miles, it again 
enters the same great valley on the opposite side of the lake, toward 
which it once more flows, this time in a north-easterly direction, to a 
point 15 miles from the lake, where it turns south. The valley, which is 
everywhere a prominent feature in the topography, is occupied by six 
small streams beside the lake and the river, three flowing west and 
three east, two to the lake and four to the river ; and in it, moreover, is 
the service of the main branch of the Tigris river. 

The name Gyuljuk means «• little lake," but the sheet of water is 
12 miles long and from 2 to 3 miles wide. Previous to 1878 or 1879 it 
had no visible outlet.* For many years its borax-bearing waters had 
been steadily rising, and at that time overflowed to the Tigris. The 
older villagers, who live along the shores of the lake, say that they can 
remember a time when the water was 20 or 30 feet lower than it now is. 
Their fathers told them that in their young days it was possible to wade 
to an island which is now separated from the shore by water 75 feet deep. 
They had heard that still earlier the island was connected with the 
mainland, and was the site of the village which now lies on the main- 
land opposite. This is incontestably corroborated by the fact that on 
the island are the remains of an old monastery,! all around which may 
be seen the ruins of houses submerged in the lake to a depth of 25 or 
30 feet. Part of the stones of the monastery are of a kind not now found 

* H. F. Tozer, * Turkish Armenia, and Eastern Asia Minor.' London, 1881, 
pp. 239-246. 

f Verhandlungen der Btrliner atdhropologi»chen Oesellscha/t, February 17, 1900, 
pp. 144, 150-152. 

Digitized by 




Castle $J 

TKe Great Canon 

of the 


By Ellsworth Huntingdon 

Scale of Miles 

» 30 mj 

•• * >i"^ j* — " .. > ii ~ I h ^— ii — 

Nat.scmlc |: 1,500.000 or 23*7 mflea « [ Inch 

Heights in feet . Routes «...-*- 

Digitized by 



in the vicinity, but probably cropping out under the lake. Some 
villagers who know how to read say that in their church was an old 
book, unfortunately destroyed during the events of 1895, in which it 
was recorded that six hundred years ago the site of the lake was a plain, 
through the middle of which flowed a stream emptying into a hole at 
the lower end. This hole was later filled with silt, and the stream was 
thus dammed so as to form a lake. In 1899 I mapped the lake and 
took a number of soundings, from which the contour of the bottom was 
roughly determined. It is perhaps worth noting that at the east end, 
just where the villagers locate the ancient outlet, there is a very 
precipitous slope with a descent of 400 feet in 1100 feet. Another 
corroborative piece of evidence is furnished by two castles which were 
built between one and two thousand years ago, and stand on opposite 
sides of the lake, near the middle. Such castles are always built 
with a definite purpose, but under the present circumstances it is 
hard to imagine what use they could serve, as there is nothing 
for them to defend. If there were no lake, the shortest and easiest 
road from Harput to Diarbeker, as determined by the location of 
passes, would go directly across the lake-basin from one castle to 
the other. Hence they lend additional certainty to our conclusion that 
no lake, or only a very small lake, existed here one or two thousand 
years ago. 

Accepting this conclusion, it is clear that the present shore-line is 
along an old beach. In twenty or twenty-five years a small lake like 
Gyuljuk would be utterly unable to have much effect in producing a 
bench along its margin. The beach as now seen, however, consisting of 
sand and pebbles, is fairly continuous, although broken in many places 
by projecting cliffs, especially on the north side, where the shale dips 
away from the lake ; and in some parts the beach is 40 or 50 feet wide, 
and is backed by a second line of wave-worn material lying 6 or 8 feet 
above the lake-level, which has been lowered by a trench dug twenty 
years ago at the outlet. Behind this upper beach are several small 
lagoons. While discussing the beach, it may be said that all the little 
brooks which enter the lake have fan deltas of coarse pebbles and 
boulders. A few of the deltas are nearly 1000 feet wide. The con- 
dition of the shore shows that this is not the first time that the lake has 
stood at the present level. Accordingly we conclude that in prehistoric 
times the lake was the same size as now. Somehow it was drained 
wholly, or in large measure, and so in early historic times a plain 
occupied most of the present lake-bed. At last, less than a thousand 
years ago, the lake-basin began to be filled, and in 1878 the water 
overflowed into the Tigris river. The traditional cause for this emptying 
and filling seems the only adequate one. Violent or extensive earth- 
movements are utterly out of the question, because there is not the 
slightest sign of them, the present shore-line agreeing exactly with the 

Digitized by 




old one. Changes in climate are impossible as an explanation, because 
a reduction in rainfall sufficient to cause the lake to shrink to the size 
which it must have assumed, would have made the surrounding country 
a desert at the beginning of our era, which, as every one knows, is 
contrary to the facts of history. Therefore we seem to be justified in 
assuming that somehow a passage was opened in the bottom of the 
lake; most of the water was drained off; later the hole was filled, 
probably by the deposition of silt, and so the lake was restored to its 
former condition. 

North of the Taurus mountains is a great lozenge-shaped basin, from 
the two ends of which flow the Murad Su and Tokma Su in opposite 


directions towards the centre. This is floored with a series of plains 
formed of a fine alluvial deposit and broken by a number of small 
mountain ridges. Cretaceous strata are abundant in these ridge?, but 
their lower layers contain more sandstone and conglomerate than do the 
corresponding strata further south. North of the plains lies the southern 
half of the anti-Taurus mountains, in which may be included the great 
Denim range. The old trap, the Cretaceous strata, and the later lava- 
flows are well represented, and there is a large development of lime- 
stone, seemingly older than the Cretaceous, although its age has not 
been determined. This is especially abundant among the highest 
mountains, those around which the Kara Su flows, where it turns from 

Digitized by 



a westerly to a southerly coarse, and in the lower mountains farther 
west. It reaches its maximum development in Muzur Dagh, where the 
thickness from the river at Egin to the top of the mountain appears to 
be 8000 feet. Between the two great divisions of the anti-Taurus 
mountains lies another series of alluvial plains at an elevation of from 
4000 to 6000 feet. They are to Asia Minor what Tibet is to Asia 
Major. Through them flow the Kara Su from the east and the Chalta Su 
from the west, meeting a few miles above Egin. The former flows on 
the surface of the larger plains, but on the sides of the latter river the 
plains lie at a much higher elevation, and have been cut by the river to 
a depth of 2000 feet. 

Such, in brief, is the country where the Euphrates is formed by the 
junction of two large streams, and changes its course from eastward to 
southward. The main roads have been often traversed and described by 
travellers, but most of the region is very imperfectly known. The great 
general Yon Moltke was the first to investigate the course of the 
Euphrates river, as it zigzags through the mountains.* In July, 1838, 
when the water was at its lowest, he floated down on a raft from Palu 
to Birejik, finding the first 125 miles to Kemur Khan easy, but farther 
down, in the gorge through the Taurus mountains, meeting with the 
greatest difficulty because of rapids. A second attempt in the spring 
of the next year, when the water was high and the rapids more 
dangerous, had to be abandoned at Tilek near the beginning of the 
greatest canon. Since that time no one, so far as I can learn, 
either native or foreigner, had attempted the journey through the 
lower gorge until in the spring of 1901 it was my good fortune to 
accomplish it in company with Prof. Thomas H. Norton, U.S. Consul at 

Starting from that place, we rode to Akhor, on the bank of the 
Euphrates, at the eastern end of the Harput plain. The people of this 
village, Armenians, make a business during the winter of floating 
down the river to Kemur Khan on rafts of skins, fishing as they go. 
These rafts are known as kelleks, and the raftsmen as kellekjis. 
As there are no equivalent English words, I shall employ the 
Turkish terms. At Kemur Khan the kelleks are taken to pieces, 
and together with the fish loaded on donkeys sent by land across the 
neck of the river's bend to meet them. The fish are sold at Harput, 
and the rafts are taken back to the village, whence they start again. 
We had engaged two of these fishermen to take us down the river as far 
as we should choose to go, with the condition, imposed by them, that 
they should be allowed to make a portage around one dangerous rapid, 
of which they had heard, in the lower gorge below Kemur Khan. We 

* See Moltke, H. von, * Briefe uber Zustande und Begobenheiten in der Tiirkei au 
den Jahren 1835 bia 1839/ Berlin, 1876, pp. 289-291, 360-363. 

Digitized by 




sent our horses by land across the narrowest part of the bend of the 
river to Chunkush, and thenoe to Gerger, and on the morning of Friday, 
April 12, were ready to begin our voyage. 

The making of the kellek took some time, although in the evening a 
number of entire sheepskins had been well soaked and left wet so that 
they might be pliable and ready for immediate use. In the morning 
they were inflated by blowing through the necks, the legs being securely 
tied so that no air could escape. At first the mouths of the blowers were 
at a distance of 8 or 10 inches from the necks of the skins, but as the 
latter became fuller and more difficult to inflate, the men's mouths were 
brought nearer until they touched the skins. When a hole was 


discovered, it wa9 quickly mended by putting a piece of wood like a 
checker against the inside of the hole and tying tbe skin firmly 
around it. A light frame of saplings was tied together with ropes, and 
under this were tied the skins, about thirty in number, with the legs 
up. They were packed together so closely as to make the kellek water- 
tight. Thirty skins seemed to us very few for five people, but the 
fishermen's rafts consist of only six, and two men sit on one such 
kellek. The kelleks always go in pairs on long fishing-trips. 

As the spring of 1901 in Turkey was unusually dry, the river was 
oomparatively low, being about halfway between the extremes of flood 
and low water. As it was, the current seemed very swift even in the 

Digitized by 



plain at Akhor. As soon as we began to float, we concluded that a 
kellek moves in the easiest, most delightful way that can be imagined. 
There is no jar or shake. The buoyant skins and pliant saplings adapt 
themselves to every movement of the waves. Half an honr after start- 
ing, we stopped for some time while the kellekjis gathered a great 
quantity of weeds, which they spread over the raft, partly to protect 
the skins from injary by onr feet, bnt still more to prevent them from 
drying in the hot sun and cracking. Every hour or two they threw 
water over all exposed portions of the skins. 

On the plain half a mile from this place, near the village of 
Eliraellik, is one of the many artificial mounds which dot the plains of 
this region in large numbers, and which contain polished stone 
hammers, bone knives, and other implements. This mound is small, 
only 200 or 300 feet in diameter, and insignificant except for 
one thing. Close beside it flows a mill-race, which has undermined 
one side so that a perpendicular section is exposed about 30 feet 
high. About 12 feet above the level of the plain is a horizontal 
stratum of water-rolled gravel, 2 or 3 inches thick. Above and 
below this the mound is composed of loam filled with bits of pottery. 
Clearly this mound was built to a height of 12 feet, and then sub- 
merged under some body of water long enough for the layer of gravel, 
and probably for some layers of finer material, to be deposited. The 
water then retired, and men again occupied the mound. The cause of 
this submergence is not clear. The first thought is that it was due to 
a river flood of unusual height, but under present conditions this would 
be impossible, for the highest level of the water is now 30 feet below 
the mound, and the range from high to low water is not more than 15 
feet. Moreover, if a flood reached to the mound it would not deposit 
gravel, but sand or finer material. 

A short distance below this mound the river enters a gorge, and a 
series of small rapids is formed. It here turns north by west, and 
passes transversely through the small Harput range of mountains. The 
gorge, cut almost wholly in basalt, has walls 1500 feet high on the 
south side, and 2500 feet on the north, the distance between the two 
summits being but little more than 3 miles. Halfway between the 
entrance to the gorge and the mouth of the Muzur Su, an imposing 
basaltic rock rises in the middle of the river to a height of 60 feet. 
On the side down-stream, where alone it is possible to land, is an 
artificial platform, from which a flight of rock-hewn steps leads to 
the levelled top. A few other places where ascent might be possible 
have been protected by walls. On the light bank of the river 
close by is a cliff, in which has been cut a cave about 20 feet deep. 
From the back of this ascends a flight of seventeen steep steps, leading 
to an opening which commands a fine view up the river. The style of 
work shows that this was a fortress of the Haldis, or Nairi, those 

Digitized by 




unconquerable enemies of the Assyrians of whom we have lately learned 
so much through the labours of Drs. Belk and Lehmann, especially the 
latter. The location of a castle here, where there could be no bridge, 
and where the road along the river-bank is very difficult and much 
longer than over the mountains, and the fact that the watch-cave 
faces up, not down, the river, indicate that it was built to guard against 
enemies who came down the river itself. Probably they used to float 
down the river on kelleks 3000 years ago just as they do now. 


The next point of interest below the castle is the mouth of the 
Muzur river, a large and very swift stream with remarkably clear 
and cool water. Bising in the mountain-girt plain of Ovajuk, and 
fed by the springs of the great Dersim mountains, from the north 
side of which the snow never disappears, it always has a full stream 
navigable for kelleks from the plain to its mouth. Its largest braoch, 
the Peri So, 150 miles long, rises near Bingyul Dagh, south of 
Erzeruna, and in spring is larger than the main stream. Thirty miles 

Digitized by 



above its mouth, near the old ruined town of Baghin,* with its Haldi 
fortress and two cuneiform inscriptions, it passes through a gorge on 
the two sides of which hot springs gush out in numerous places. For 
a quarter of a mile the river is confined between perpendicular walls 
covered with stalactites 40 or 50 feet long, the deposit of the springs. 

Passing the mouth of the Muzur Su, we soon came to a ferry owned 
by one of the few beys, or feudal lords, whose power is still absolute. 
His Kizilbash retainers, a few days before our visit, had stolen the 
ferry-boat from Akhor, our raftsmen's village, in order to use it in 
ferrying wood to be sold in Harput. The mountains south of the 
river are wholly deforested and the Kurds are rapidly cutting the 
small growth that remains in Dersim. The men at the ferry got out 
their old flintlock guns and tried to force us to pay for floating past 
the place where they had the ferry rights, but our hats overawed 

Below this the river leaves the mountains, but the valley is still 
hemmed in on the south for some miles by a wall from 500 to 800 feet 
high, capped with from 50 to 100 feet of basalt, the edge of a lava-flow 
which came from the mountains north-east of Harput and reached the 
edge of the river-valley, but apparently did not fill it. From here the 
river flows for 30 miles through low hills and plains broken only at 
Pertag. The first village passed on the right is Till, which must once 
have been an ecclesiastical centre of some importance, since there are 
the remains of seven Syrian churches, beside those of baths and houses. 
Indeed, the whole of Dersim is fall of relics of the Christian population, 
Armenian and Syrian, that once filled it. At Peftag, close to Till on 
the same side, ruins of other kinds are found. From far up the river 
the picturesque castle is seen on its high pointed rocks. It was first 
built by the Haldi and last rebuilt by the Seljuk Turks, but no one 
knows by how many races it has been occupied between these extremes. 
At the base of the castle are the ruins of several mosques, churches, 
and baths, and of hundreds of houses. At the beginning of this century 
Pertag seems to have been a large and flourishing town, but in 1 839-40 
troops were brought here and quartered in the houses of the people, 
who promptly moved to their beautiful well-watered gardens, an hour 
away to the north-east at the base of the mountains. When the soldiers 
finally departed, the people did not care to return to their injured 
houses. Some ferrymen and a few zaptiehs, gens d'armes, are the only 
inhabitants of the old town. A fair is held here every Friday morning, 
and is attended by Christians and Mohammedans from all sides. 

The river here cuts through a great boss of granite porphyry which 
rises in the sharp castle rock on one side and in a magnificent dark 

* Verhandlungen der Berliner anthropologUchen GetelUehaft, Noyember 18, 1901, 
pp. 174-180. 

Digitized by 



cone, 1000 feet high, on the other. Along the river the rook forme 
fine though small columnar palisades. Just west of the castle, where 
the river leaves the palisades and enters a region of black basalt, we 
discovered the ruins of an old Roman bridge, which formed part of the 
road from Harput, through Kuttu Dere, in the midst of Dersim, to 
Erzinjan. This route vid Erzinjan is the shortest from Harput to the 
Black Sea at Trebizond, but it is not now used, because the Kurds in the 
mountains through which it passes make travelling very unsafe. 
The makers of the bridge utilized as piers two islands, on the larger 
of which massive limestone blocks still remain in plaoe. The rest of 
the way to the Kara Su the scenery is uninteresting, the open valley 
being cut through limestone, which forms low rounded hills a few 
miles back from the river. On the right bank one of the villages is 
Kogpenig, where many of the inhabitants live in caves. At Ashvan 
a model farm is being conducted on American principles under the 
direction of Prof. Norton. 

The mouth of the Kara Su is disappointing because its appearance 
is not equal to its geographical importance. Both rivers flow slowly 
here, and approach each other from nearly opposite directions — a most 
peculiar configuration for well-established rivers of such size. In the 
summer of 1901, I floated down the Kara Su from Egin to the Malatia 
plain. The canon at Egin is one of the finest in Turkey. Two miles 
above the town the narrow stream flows between solid walls of hard 
limestone 400 feet high, which even, when looked at from a distance, 
appear to be really perpendicular. Above these perpendicular walls 
the steep rough limestone rises 4000 feet on the west side in a distance 
of only 4 miles, and on the east side 8000 feet in scarcely 8 miles. 
Trees and vegetation are almost laoking, and the landsoape is all brown 
and grey; yet, in spite of the bareness, it is grand. Egin itself, 
thanks to the great springs, is completely hidden in trees, so that the 
contrast between the green city and the bare mountains is most 
striking. In floating down the stream, the mountains are soon left 
behind, and the old limestone falls lower and lower, until at last it 
forms a wall but 10 feet high. Over this lie sandstone and much rough 
conglomerate, whioh seem to be a part of the great Cretaceous deposits. 
On both sides of the river the land rises 1000 feet or more in broad 
irregular terraces to a partly dissected plain covered with alluvium. 
Near the mouth of the Kara Su, the old limestone rises again in 
rounded hills from 400 to 1000 feet high. On the Kara Su below 
Erzinjan only one dangerous rapid occurs, 15 miles below Egin, and 
it is dangerous only because of the great number of stones in the 
middle of the channel It is worthy of note, as an indication of the age 
of these streams, that, although the grade is in some parts steep, it is 
rarely broken by ledges and sudden descents. Except in the canon 
between Kemur Khan and Chunkush, the rapids of both the Kara Su 

No. II.— August, 1902.] o 

Digitized by 



and Murad Su are usually due to accumulations of gravel brought 
in by side streams. 

The mouth of the Kara Su lies at the head of a second winding 
transverse gorge cut through limestone except at Geban Mad en, where 
there is an outcrop of trap in which are silver mines. The raftsmen 
told us that 2 miles above Maden there were traces of another bridge, 
which, however, proved to be a quarry with a series of flat steps full of 
drill-holes, and cut into the face of the steep cliff in such a way that 
blocks of the beautiful variegated marble could be loaded on to rafts 
with the greatest ease. Where the marble was carried is not evident, 
as none of it has been noted in neighbouring ruins. The passage 
through the real gorge, beginning at Maden, occupies four hours 
with a swift current, and in spring some good rapids, which are 
scarcely noticeable in the lower water of summer. The limestone walls 
tower very steeply 1000 feet or more, and above that height the 
mountains rise another 1000 or 2000 feet. Hundreds of pinnacles and 
peaks rise like countless castles separated by gigantic clefts. Here a little 
tributary comes at the same level through a canon with perpendicular 
walls ; there a sheer cliff rises 500 feet ; close by, the massive strata are 
crumpled like paper or are set at various angles by great faults. 
Where the strata are horizontal, the wall is benched with terraces from 
20 to 60 feet high, each bearing on its top a strip of beautiful green 
grass in delightful contrast to the prevailing buff grey of the mountains 
and intense blue of the sky. Almost the only inhabitants are big-horned 
ibex and wild blue pigeons, which mako their home in the numberless 
inaccessible caves which honeycomb the limestone from top to bottom. 
In the intense heat of summer, when everything dries up and the 
canon is like a furnace, even these are not seen. Near the lower or 
south-west end of the gorge the walls grow steeper as they decrease in 
height, until the river passes out into the Malatia plain from between 
perpendicular limestone walls, here only 40 or 50 feet high. 

The Euphrates, now turning south, skirts the base of the western 
extension of the Harput mountains, entering but slightly the great 
Malatia plain which stretches 20 miles to the west. The inhabitants on 
both sides are for the most part Kurds, those on the left of the river 
being largely Zaza, those on the right, north of the Euru Chai 
(Chai= brook), being Eizilbash, and those on the right, south of the 
Tokma Su, Kurman ; while the area between the Kuru Chai and the 
Tokma Su, very fertile, but not easily irrigated, is practically un- 
inhabited. The Eizilbash are the most interesting of these three 
divisions of the Kurds. They are a mixed race, the foundation 
being some tribes of a stock allied to the Persians, who advanced 
into Turkey along the central highlands. These mountains were 
inhabited by Armenian Christians,who under stress of persecution became 
nominal Mohammedans and intermarried with the invaders. The 

Digitized by 



Kizilbash in the district near Mala tia, unlike their brethren in Dersim, 
are peaoeable, well-behaved agriculturalists, most of whom have entirely 
given up nomadic life. In general the Kizilbash are a frank, good- 
natured people, eager to be amused, very ready to ask and answer 
questions, hospitable, easy to deal with, industrious when work is 
necessary, and faithful when they have given their word, although very 
ready to rob and even to kill those to whom they are not under obligation. 
Morally they are superior to their neighbours. They deteriorate rapidly 
under new or adverse conditions, becoming more suspicious and 
treacherous. When among the Turks, they swear that they are good 
Sunni Mohammedans, although in reality their religion is a mixture of 


Shiite Mohammedanism and Christianity, with perhaps a trace of 
primitive paganism. Accurate information is hard to obtain, because in 
talking with a Christian they try to make their religion appear like 
Christianity. For instance, a prominent agha, or village chief, said to 
me, *' We have four great prophets, Adam, Moses, David, and Jesus, of 
whom Jesus is the greatest. We have four holy books, the Gospels. All 
religions are but different roads to the same end — one long, one short — 
one easy, one hard. You go yours, and we go ours." When I tried to 
talk about Mohammed, he avoided the subject as though it were 
unpleasant, so that I could learn nothing. The Kizilbash never pray 
in private, but only when led by one of their sehids, or religious chiefs, 


Digitized by 



who have great influence among them, and who go freely and safely 
from tribe to tribe even during times of fead. So common are fends, 
and so respected are the sebids, that the business of conducting travellers 
or of carrying freight is wholly in their hands. At certain times they 
observe a sort of sacrament, which closely resembles the Christian Com- 
munion Service. No competent observer seems to have witnessed this, 
and it is known only by report The Kizilbash reverence all Christian 
sanctuaries and churches, and will even go into a church where service 
is being conducted, and take part, kneeling and bowing with the people. 
To be sure, they will do the same thing in a Sunni mosque, but in the 
latter oase it is for fear of persecution, while in the former it is a matter 
of their own inclinations. 

The Euphrates, as it winds through the Malatia plain, flows slowly 
and divides into a network of channels, enclosing islands of sand 
or gravel nearly level with the flood-plain. The latter, often half a 
mile wide, is bounded by bluffs from 30 to 50 feet high, cut in the 
alluvial deposit which forms the Malatia plain. Villages, especially on 
the left side, are numerous and prosperous, being usually beyond the 
flood-plain, although some lie at its edge. 

At Kala, close to the most western point of the great bend, there is, 
on the left side of the river, a large rock, which the Haldis, or Nairi, 
long ago fashioned into a castle by excavating platforms and steps, and 
building walls. Behind the village, on the steep slope of Mushar 
mountain, are situated several famous holy places. The first, 400 feet 
above the river, is a raised platform of stone and mud, said to be the 
grave of an Armenian girl who cared for the great church on the top of 
the mountain. The Kizilbash aghas of the village are honoured by 
being buried here, although the common people must be content to lay 
their dead by the river. Five hundred and fifty feet higher is found a 
Turkish holy place, the grave of a man called Hassan, in a small cave, 
which has been walled in and furnished with many gaudy and some 
valuable offerings. Outside is a great square altar of rough stones, all 
covered with the gore of the scores of sheep and goats, which are brought 
as sacrifices by both Christians and Mohammedans, and which are cooked 
in huge copper caldrons hung from great beams. The horns of the 
offerings are piled on another altar, and the meat is often eaten in the 
holy place itself, the bones being thrown into a little cave back of the 
main cave. The shrine has no guardian, but it is regarded with such 
veneration by men of all religions that the most valuable of the offerings 
are perfectly safe from pilferers. The third, and least visited, holy 
place is an old well-built churoh with massive buttresses and arches, 
located on the bleak mountain-top 2400 feet above the river. A more 
unpromising or inaccessible site for a churoh can hardly be imagined. 
The view from Surp Aharon, as the Armenians call the churoh, or 
Mushar Kilise, as the Turks call it, is very extensive, including the 

Digitized by 




snow-capped Dersim and Anti-Taurus mountains on the north, the 
western extension of the Harput mountains on the east, the Taurus 
mountains on the south, and on the west the broad brown expanse 
of the Malatia plain, bounded on the east by the blue network of the 
Euphrates, and on the west by the range of Aghaja Dagh, out by the 
V-shaped cleft of the Tokma gorge. 

Returning to the river, the reedy, bushy islands, or the banks of 
shingle between the branches of the river, are the resort of all kinds 


of water-birds — ibis, black divers, storks, bustards, herons, cranes, and 
many smaller birds. The current was slow and the voyage monotonous 
as we floated past the mouth of the Kuru Chai, whose valley the 
Malatia-Sivas road follows for two days' journey, and past the 
large Tokma Su from Gurun and Derende, until we reach Pirot, where 
the road crosses the Euphrates. Here the mountains again approach 
the river, leaving on either side a strip of smooth green fields dotted 
with trees and houses. Behind this on the right rises the fcrst ridge of 

Digitized by 



the Taurus, from 4000 to 5000 feet above the river, and green clear to 
the enow-line, where verdant wheatfields lie olose to snow-filled valleys 
from which shining little brooks tumble down the steep slope. Although 
the view is not wild or grand, it is most attractive, because of its quiet 
strength joined with real verdure — a rare element in the landscape of 
Eastern Turkey. 

For 6 miles we skirted the base of the mountains, whioh, approaching 
gradually nearer the river, cause it to become more narrow and swift. 
Near Eemur Ehan, on the left side, is a cuneiform inscription recount- 
ing an expedition of Tiglath Pileser, and speaking of a certain bridge, 
presumably over the Euphrates. Just up-stream from the inscription 
is a fairly narrow place in the river, with low cliffs on either side which 
might readily serve as piers of a bridge. As we passed this point, our 
kellekjis volunteered the information that they had heard from their 
fathers that in old times there was a bridge here, of which they — the 
fathers — had in their youth seen a few stones. 

At Eemur Ehan the river turns at right angles and goes south by 
east through a remarkably straight gorge 12 miles long and nearly 
4000 feet deep. The scenery is even finer than in the preceding gorges. 
The dark, steep, gloomy walls of basalt and metamorphic shale are 
terraced at an elevation a few hundred feet above the river, and on each 
terrace or nestled in each tiny valley are one or two houses and a patch 
of bright green fields. In some cases the fields are on slopes so steep 
that it seems as though the sower could scarcely find a footing. Far 
above the fields white patches of snow contrast strongly, in spring at 
least, with the black and green walls of the canon, and send little streams 
cascading down through rough gashes in the resistant rock, amidst a 
chaos of huge boulders and trees. In this gorge our real difficulties 
began. Our Armenian kellekjis, who knew the river thoroughly, as far 
as Eemur Ehan, were now beyond their accustomed track and ready to 
be afraid of everything. The first rapid in the gorge looked to them so 
bad that we made a portage of between 2 and 3 miles around both that 
and the next rapid, climbing 1200 feet up the steep slope over the 
roughest kind of road. If the kellekjis had not been so timorous, the 
rapids might easily have been shot, and we should have been saved 
twenty hours. 

As we were about to embark below the portage, after a night spent 
on the river-bank, a raft of logs passed us manned by two almost naked 
Eurds, with wooden tridents in their hands, and strings of gourds 
around their waists for life-preservers. They carry wood from Izoghlu 
through the Eemur Ehan gorge to Eefferdis. We followed the men, 
and after a few minutes shot a fairly large rapid, at the bottom of which 
was a big whirlpool. Our men, with their spoon-shaped walnut 
paddles, were able to pull us out of it, but the Eurds could do 
nothing with their tridents, and were carried far up-stream. Finally 

Digitized by 



they had to land and haul their raft down to a point below the 

Through the whole length of the gorge we went at an average rate 
of 5 miles an hour, between walls of solid rook which oome down sheer 
to the narrow stream, and are broken only by precipitous gullies 
entering at grade and bounded by jagged cliffs with needle-like points. 
The mouths of these gullies are footed by fan deltas, which have been 
pushed out into the river, forming damp, over the outer ends of which 


the water pours in foaming rapids. We shot into these over smooth 
rounded waves, like the long swells of the ocean, but in a moment were 
among the breakers, which tossed the light raft up and down like a 
oork, and often came over us, breaking up-stream, as is usual in rapids. 
The kellekjis paddled with all their might. The raft spun round and 
round, so that we saw the wild mountains on every side without turning 
our heads. 

Near the end of the gorge, where the mountains become lower, a 

Digitized by 



good-sized brook from Morjimud, coming in on the left through a fine 
oanoo, has bnilt a large fan. The size of the rapid thus formed 
frightened our kellekjis so much that we could not persuade them to 
shoot it, and were obliged to spend two hours in letting the kellek down 
past it by ropes. While we were doing this, the Kurds on the wood- 
raft overtook us. Lying flat on the logs, they shot through, not over 
the rapid, going into wave after wave with 4 or 5 feet of water over 
them, and ooming out at the bottom with a triumphant yell. 

At the mouth of Mamash Chai we left the transverse canon, and 
turned at right angles to the east by north into the broad longitudinal 
valley in line with Lake Gyuljuk. On looking into the valley from 
any of the surrounding mountains, it appears so broad and well matured 
that one feels certain at first that the Euphrates must come down its 
whole length from the west-south-west, the actual canon of the river 
seeming to be merely the bed of a short tributary. The part of the 
main stream in the open longitudinal valley and the tributary, Mamash 
Chai, lying directly in line with one another and at right angles to 
the transverse oanon, have incised in the valley-floor a steep-sided 
trench from 150 to 250 feet deep, and wholly filled at the bottom by the 
streams, which flow between rook walls without a flood-plain. Most of 
the valleys in this immediate region show the same feature of a flat- 
bottomed old valley with a small trench-like new valley, without a 
flood-plain, incised in the floor. Thirty miles north in the Harput 
mountains such newly incised valleys are not found. They occur on a 
small scale 30 miles north-west, in the western extension of the Harput 
mountains, around which flows the Euphrates river, and again 20 miles 
further east, in the valley of the Tigris river. Probably this con- 
formation is due to an uplift of the region at a comparatively recent 
date, affecting most strongly the Taurus moan tains where they are 
traversed by the Euphrates river. The minimum elevation seems to 
have been at least 500 feet. The immediate vicinity of Harput may 
have been unaffected, or, as is equally probable, the outting down of the 
bed of the Euphrates has not yet proceeded far enough up-stream to 
affect these mountains, although they may have been uplifted. In the 
Tigris basin the very small dimensions of the newly incised valleys 
may be due to slight uplift or to the comparative slowness with which 
the small stream of the upper Tigris scours out its bed. The 8 miles 
of the course of the Euphrates, along the open longitudinal valley in 
line with Lake Gyuljuk, are over numerous rapids, caused not by fans 
but by inequalities in the bed-rock, and thus indicating that the revival 
of the river by uplift of the surrounding country is very recent. 

Near Aivose, just below the ferry where the road from Harput to 
Shiro crosses the river at the head of a dangerous rapid, which, unlike 
those just above it, is caused by a detrital fan, thirty or forty of the 
villagers tried to prevent us by force from going further, but were soon 

Digitized by 



persuaded to take up our oraft and carry it a quarter of a mile around 
the rapid. They could not understand why any one should make such 
a dangerous journey, and especially why any one should take notes all the 
time, and so, with oriental logic, they concluded that we had some secret 
purpose which must be opposed to their interests. Our own servants 
were equally unable to understand our purpose, although I tried to 
explain. Often I heard them answering the questions of inquisitive 
villagers, "What are these men doing? " " We don't know. Perhaps 
they are going to make a bridge, or a wagon-road, or a railroad. 
More likely they have a secret commission from the king. They say 
they are not paid for making this journey, but we know better; they 
are not such fools as all that. They know everything : they see a stone 
or a plant, a brook or a moun- 
tain, and they know it. They 
even know what is in a place 
before they have visited it. 
They write everything. If you 
want to know any more, ask 
them. What do we know ? " 

The villagers at Aivose 
feared not only that we should 
harm them, but also that we 
should be drowned, in which 
case the Government might 
hold them responsible for the 


foreigners, the men with n inflated sheepskin. 

hats." The people here, as 

well as many others with whom we talked, both above and below this 
point, asserted that no one ever had navigated or could navigate the 
river from Aivose to Chunkush. Von Moltke's journey seems to have 
been forgotten. 

Leaving the villagers, who had become quite friendly as they 
carried our goods around the rapid, we boarded the kellek once more, 
and in less than ten minutes were at the angle where the river, turning 
onoe more to the south from the longitudinal valley, enters an immense 
crooked transverse canon, the last and longest of the great gorges, 
30 miles long and 5000 feet deep. Before we knew it we were at the 
head of a rapid worse than any that we had yet shot, or around whioh 
we had made portages. It seems to be due partly to the structure of 
the bed-rock and partly to the fan of the Uslu brook, which flows into 
the river just in the middle of the rapid over a series of small cascades, 
which, as seen from the river, appear to be caused by the brook's own 
fan. The kellekjis wanted to make another portage, but we insisted on 
shooting the rapid. Although we made the passage safely, the men's 
nerves were bo completely unstrung that when we landed soon after at 

Digitized by 



the bead of another large bed-rock rapid, they absolutely refused to go on. 
One was sent to find a village and get men and animals to help in 
making a portage. He found a village after a hard climb of an hoar or 
more, but instead of going to it he bid till after dark, and then returned 
to us, because, as he said, if the Kurds knew that we were encamped 
beside the river, they would come and rob or even kill us during the 
night. The other Armenian, when told to take some baggage off the 
safely moored kellek, said, " If I ever set foot on that kellek again I 
know that I shall die. Then who will take care of my wife and 
children? You haven't any hearts. The mountains are savage, the 
river is savage, the people are savage, but you don't fear them. Don't 
you even fear God ? " 

He was full of superstition and was in terror, partly because he had 
dreamed the night before that some of his friends were dying. This 
night he dreamed that we Americans were shooting pistols, which, 
curiously enough, seemed to encourage him greatly. 

In the morning we made our portage, a long hard one involving a 
climb of 900 feet up out of the steepest parts of the canon, and at twenty 
minutes past five in the afternoon were ready to embark, 2 miles 
from our camp of the previous night. Floating very rapidly for nearly 
an hour, we passed the sulphurous hot springs of Tilek, which rise on 
both sides of the river below high-water mark. They contain chiefly 
hydrogen sulphide and calcite. In July, 1900, I visited them and 
found the temperature to be 114° Fahr., while that of the surrounding 
air was 103° Fahr. at 4.30 p.m. On account of the inaccessibility of the 
springs, they are but little visited, although reputed very beneficial 
for skin and rheumatio diseases. The favourite method of treatment 
is to bathe the patient half an hour, and then bury him up to the 
neck in the hot river sand for two hours, repeating this process four 
or five times in a day. 

Near Tilek the river turns more directly south and passes through 
magnificent scenery. At the bottom, schist, formed by the metamor- 
phism of shale, forms almost perpendicular walls, which we estimated 
as nearly 400 feet high. Above is a terrace, from which green wooded 
upper walls rise less steeply to the mountain-tops a mile above our 
heads. The larger tributaries cut through the apparently perpendicular 
lower wall in steep-sided narrow canons, entering the main stream at 
grade, while several small streams have not yet cut down to grade, 
and cascade over the cliffs in a series of white falls. 

These hanging valleys and many other characteristic features of 
the Euphrates canon correspond closely to what is described in the 
Grand canon of the Colorado river, showing that the two rivers are of 
nearly the same age. The Colorado canon is much longer than that of 
the Euphrates, and is cut through a plateau of nearly horizontal strata 
instead of through a range of mountains formed of muoh distorted 

Digitized by 




strata ; but in both cases the depth of the canon and the inclination of 
the walls, very steep at the bottom and more gentle above, are approxi- 
mately the same. Both rivers are so young that they have not yet 
established thoroughly graded channels, but no waterfalls are found 
each as we should expect in extremely young streams.* There are, 
however, the two kinds of rapids which I have already mentioned, viz. 
those due to outcrops of hard rock which have been worn back so far 
as to present a regular but nevertheless very rapid descent, and those 
due to the damming of the stream by fans of detritus brought in by 
tributaries. It was by these that the famous journey of Powell down 


the Colorado was made so dangerous.f We found that as a rule the 
bed-rock rapids are longer, but the fan rapids are more steep and 
dangerous. They indicate that the rivers are approaching, but have 
not yet reached, the stage of maturity, when the whole channel assumes 
a graded charaoter. 

The junction of the side streams with the main river is another 
interesting indication of youth. The majority, as has been said, enter 

* See Gilbert, G. K., Engineer Dept. U.S. Army, ' Report upon Geog. and Geol. 
Expl. West of 100th Meridian,' vol. iii. (Washington, 1875) part i. pp. 70-75. 

t See Powell, J. W., « Exploration of the Colorado River of the West and ita 
Tributaried.' Washington : 1875. 

Digitized by 



at grade through narrow steep-sided canons, but some of the smaller 
tributaries and a number of wet- weather torrents have not been able to 
carve valleys as rapidly as the main stream has incised its canon, and so 
have been left hanging at various levels on the canon wall. They are 
in marked contrast to the hanging valleys desoribed in glaciated coun- 
tries such as Switzerland and Norway. The latter enter main valleys 
far broader than the streams that occupy them — so broad, indeed, that the 
main rivers have room to wander over wide flood-plains, bordered often by 
narrow strips of plain which graduate into the steep valley wall with- 
out any sudden angle. The others, on the contrary, send their streams 
cascading over the precipice directly into the Euphrates river, whose 
valley is so narrow that not only is there no flood-plain, but the water 
washes the rocky base of the almost perpendicular cliffs. The Colorado 
river, it will be remembered, has just such hanging valleys where it 
passes through the harder crystalline rocks. These two oases, where 
the discordance of side valleys seems to be incontestably due wholly to 
river-erosion, are in suoh marked contrast to the discordance of valleys 
of glaciated countries, that it is very hard to believe that the latter 
could have been due to river-erosion. If we regard them as the product 
of glacial erosion, we have what seems to be an adequate explanation of 
their difference from those that are certainly of river origin.* 

In the middle of the fan delta of Haloge Chai, a small tributary 
of the Euphrates which enters the main stream at grade, lies an immense 
white boulder, 30 feet in diameter, on the top of which are perched a 
number of boulders of other kinds, small only in comparison with the 
monster on which they rest. Probably the surface of the delta was 
formerly high enough to cover completely the large boulder above which 
the others were laid down. As the delta was worn away, the latter 
were left resting on the former. That water could transport and smooth 
a boulder of such enormous size seems incredible, but the stone is there 
at a considerable distance from its point of origin, and must have been 
carried down the slope to its present position by floods.f 

A mile below this we stopped at the head of the biggest rapid yet 
seen, and of course the men wanted to make a portage. As our time 
was growing short, and a portage would take nearly a day, we decided 
to leave the light baggage for the men to carry over the quarter-mile 
chord of the great boulder-strewn fan which caused the rapid, and ourselves 
take the raft down. They expected that we should be drowned, and 
they would be left alone without a kellek in the bottom of that almost 
inaccessible gorge ; but in spite of their entreaties we pushed off, and 

♦ See Davis, W. M., «• An Excursion to the Grand Cation of the Colorado," Bulletin 
of the Museum of Comp. Zoology at Harvard College, vo 1 . xxxviii. (Cambridge, 1901) 
p. 169. 

f See Gilbert, G. K., op. cit. y p. 71. 

Digitized by 




although we paddled hard were in the rapid before we had reached mid- 
stream. There was a long swift exhilarating shoot over a tilting stretoh 
of water, and we were among dashing waves which seemed to be 10 feet 
high. As the kellek rose on the first one, we stopped paddling and 
seized the ropes. We whirled round and round more swiftly than ever 
before, this time not looking at the grandeur of the mountains, but 
only at the waves, whioh broke over us again and again, wetting us to 
the skin. The kellek stood the passage perfectly, and below the rapids 



we brought it safely ashore. When at last the men overtook us, they 
seemed to feel that our preservation was miraculous. 

" You're not men; you're jinner (spirits) ! " one of them exolaimed. 

From here to Chunkush the gorge is cut almost entirely through 
limestone. Travelling at the rate of 8 miles an hour, we passed for 18 
miles through a continuous succession of bed-rock rapids, many of them 
larger than those around whioh we had made portages. We spent the 
night on a tiny ledge, where the precipice overhung us and partly 

Digitized by 



protected us from the light rain that fell toward morning. Although 
neither the river, the mountains, nor the Kurds had harmed us, the men 
seemed determined to find something of whioh to be afraid. 

" If we sleep here, the bears will come to the top of the precipice and 
throw stones on us," they said, when we proposed to land on the narrow 

The next to the worst rapid is 2 miles above Chunkush, where the 
one of our kellekjis whose sleep had been disturbed by dreams was washed 
overboard by the waves. Most fortunately his companion seized his arm, 
and we all pulled him on to the raft at the imminent risk of upsetting it. 
A few minutes later, while we were still in the same long rapid, we 
encountered one of the dangers that we had most feared. The raft 
stuck on a hidden rock in such a position that the waves would soon 
have battered the skins till they .leaked and the raft went to pieces ; but 
we got off by shifting the load first to one side and then the other, thus, 
as it were, prying the raft over the stone. 

From Chunkush to Gerger, where our river journey ended, the 
mountains grew gradually lower. The river presents no special obsta- 
cles and no points of marked interest, although the soenery is still fine. 
The large Chermug Chai, a brook which enters on the left, has two main 
branohes : one, the Eizilohabuk, is a wet-weather stream from the 
south-east, where it drains part of the plain north-west of Diarbekir ; 
the other has always a large stream. Its main source is halfway 
between Chunkush and Chermug, where there is in the limestone a 
great sink-hole over 400 feet deep, with an aperture scarcely 50 feet by 
20. The water from this reaches the surface 2£ miles farther east, near 
the village of Sinek, from which the stream takes its name. A smaller 
branch comes from the locally famous sulphur hot springs of Chermug. 

At Gerger, where the Euphrates passes through the last outlying 
ridge of the Taurus mountains, we find one of the narrowest gorges. 
The hard limestone mountains rise over 2000 feet above the river, 
the north side of the ridge being bounded by a long line of cliffs 400 or 
500 feet high, from tbe base of which there is a much more gentle slope 
to Fetterge Chai. The south side slopes off gradually in well-rounded 
gentle hills which can be traversed with great difficulty because of the 
rough pits and sharp edges into which the limestone weathers, and whioh 
have been denuded of soil because of the deforesting which has taken 
place. As one stands on the top of the ridge, a long broad depression is 
here visible — the old valley of the Euphrates stretching to the south- 
west nearly parallel to the mountains — but there is no sign of the river, 
whose present course is marked only by a slight line of cliffs. An 
observer who did not know that the river intervened would think that 
he could easily walk to the hills on the opposite side. The river now 
flows in a narrow canon with perpendicular walls 450 feet high. 

This journey of 190 miles by water from Akhor to Gerger occupied 

Digitized by 



seven days, including the unnecessary portages, although the time 
actually required for floating down the river was only thirty-seven hours* 
The 50 miles from Akhor to the Kara Su occupied teji hours, the 35 
miles to the beginning of the Malatia Plain six hours, the 45 miles to 
Kemur Khan ten and a half hours, and the remaining 60 miles to Gerger 
ten and a half hours. The total descent as measured by aneroid is 1250 
feet, the steepest part being near the big rapid just below Tilek, 10O 
feet in 6 miles, and the gentlest descent being in the Malatia plain, 100 
feet in 55 miles. We found that the map as given by Von Moltke needs 
considerable correction. 

At present the roads avoid the gorges of the river and climb over the 
high mountains. As soon as railroads are introduced, there can be no 
doubt but that, as usual, one of the main lines will follow the easy grade 
of the river. 

Although the utilization of the Euphrates river for industrial 
purposes may be far in the future, it is, nevertheless, a problem which 
will some day be of great importance. The most valuable part of the 
river's basin is Mesopotamia, which begins south of our journey's limit, and 
of which I cannot speak from personal knowledge. A comparison of 
the accounts of ancient and modern travellers shows, however, that its 
present poverty is almost as remarkable as was its ancient fertility and 
wealth. That region is described as being so rapidly and completely 
ruined by the invasion of the sand of the desert on the south-west side, 
and by the constant shifting of the channel of the Euphrates river in 
the centre, that the future will be even more desolate than the present 
unless measures are taken to resist these encroachments. Against both 
the river and the desert the means of protection lies in controlling the 
river itself. In ancient times the control of the water was effected by 
a great system of canals, embankments, and reservoirs in Lower Mesopo- 
tamia. These were in constant need of expensive repairs, and their 
breaking was the cause of frequent disaster. England has found that 
the only way to control the Nile is to control its sources, and, having 
finished the work of cutting the sudd, she is now making the huge dam 
at ABSuan, and considering the far greater task of regulating the out- 
flow of the great African lakes. It is a work of great magnitude, in- 
volving an immense outlay of capital, but there is no doubt that it will 
prove permanently successful. The problem presented by the Nile is 
similar to that presented by the Euphrates, and the two must be similarly 
dealt with, although there is one important difference. In the former 
case, the chief problem is to conduct the water to the sea by the right 
channels, in the right quantities, and at the right time. In the 
Euphrates river, the problem is to use the water before it reaches the sea. 

When the Euphrates river is properly controlled, it will serve two 
great uses : it will be a great producer of power, and it will accomplish 
the vastly more important work of irrigating Mesopotamia. More than 

Digitized by 



three- fourths of the rainfall of the upper Euphrates basin falls in the 
seven months from October to April. If the precipitation of this 
period could be retained among the mountains, and allowed to escape 
during the late spring and summer, it would not only save Mesopotamia 
from the annual floods whioh carry away great sections of land along 
the banks and convert other large areas into disease-germinating 
swamps, but would also accomplish the far more important result of 
enabling vast tracts of the best kind of land to be irrigated and re- 
claimed from the desert The extent of the entire region that can be 
reclaimed by means of the river seems to be as great as that of England. 
Although Mesopotamia can produce two crops a year, its population is 
scarcely a million ; yet it might support ten times as many people, and 
still export in great quantities all sorts of food-stuffs. It is destined to 
be one of the richest parts of the world if its development is unhindered. 
The great need, so far as merely physical matters are concerned, is the 
controlling of the Euphrates river. 

The best means of accomplishing this seems to be by building 
reservoirs in the great depression of the upper Euphrates valley between 
the Taurus and anti-Taurus ranges. This oontains a series of small 
parallel mountain ridges enclosing many plains, generally, although on 
slight ground, called lake-basins. In the middle of each is a stream, 
which passes out at the lower end through a steep narrow valley. The 
plain b can easily be converted into reservoirs by the construction of 
simple dams. The expense, which would not be excessive, would 
certainly be met, iu part at least, by the water-power thus rendered 
available, while the storing of the surplus water of the winter would 
be of incalculable value in irrigating Mesopotamia. Of course, the 
value of the land covered by the reservoirs must be considered. Fortu- 
nately it is slight, for although the larger plains are very valuable, the 
smaller ones are often so cut up by small watercourses that they have 
little value for agriculture, and might readily be converted into lakes. 
It is the work of the engineer to determine the best sites for dams, but 
the possibility of building them seems patent to any one who has seen 
the plains and gorges. The progress of civilization will necessitate the 
building of such reservoirs, and the only place where they can be 
profitably built is north of the Taurus mountains. 

What has been said of the Euphrates river is, with certain modifica- 
tions, true of the Tigris also. Reservoirs can probably be built east of 
or among the Zagros mountains, just as they certainly can be built 
north of the Taurus mountains. From the earliest times until now the 
history of the Mesopotamian plain has been sharply separated from that 
of the mountains to the north and north-east. In the future the history 
of the two regions must be one, because the development of Mesopotamia 
depends absolutely on the great rivers which flow from the mountains, 
and must there be controlled. 

Digitized by